Vehicle Image Transfer Using Map-Based Difference Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in efficiently communicating and storing large volumes of image data, which can be costly and difficult due to the high volume of acquired image data, necessitating a method to reduce data transmission and storage requirements.
Innovation Solution
The system uses imaging sensors to identify relevant objects in the environment, deriving a list of these objects and transferring this list instead of the full image data, allowing a remote server to synthesize an approximation of the image by rendering objects at their proper locations, and utilizes predefined maps to remove redundancies and compress difference images, reducing data volume.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If full image data is transferred from the autonomous vehicle, then complete environmental information is available, but data transmission volume and storage requirements increase significantly
Solution Approach 1:
The patent extracts only the essential semantic information (object identities, locations, and attributes) from the full image data, transferring this extracted information to the remote server instead of the complete images. This extraction principle resolves the contradiction by maintaining the necessary environmental information while dramatically reducing data transmission volume.
Solution Approach 2:
The patent introduces a remote server as an intermediary that receives transferred data, synthesizes images by rendering objects at their proper locations, and reconstructs the environmental representation. This intermediary enables efficient data transfer while preserving complete environmental information through server-side synthesis.
2Measurement precision
If detailed image data is transmitted to the remote server, then accurate environmental representation is achieved, but transmission time and latency increase
Solution Approach 1:
The patent extracts only critical object-level information (identities, positions, attributes) rather than transmitting complete detailed images. This extraction maintains sufficient accuracy for environmental representation while dramatically reducing transmission time by sending only essential data elements.
Solution Approach 2:
The patent transfers partial information (object semantics rather than full image details) that is sufficient for the remote server to reconstruct accurate environmental representations. This partial action approach achieves the necessary accuracy without the time cost of transmitting complete detailed data.
3Loss of information
If complete object lists including all background objects are transferred, then comprehensive environmental data is available, but data volume increases unnecessarily
Solution Approach 1:
The patent extracts and transfers only relevant object information from the complete scene, filtering out redundant background objects. This selective extraction maintains completeness of necessary object information while reducing data volume by excluding unnecessary elements.
Solution Approach 2:
The patent applies different treatment to different objects in the scene, transferring detailed information for relevant foreground objects while omitting or simplifying background objects. This local quality approach ensures comprehensive information for important elements while reducing overall data volume through selective detail retention.
Data Source
AI summary
A system includes at least one imaging sensor and a processor. The processor is configured to acquire detected data describing an environment of a vehicle using the at least one imaging sensor; derive reference data which describes the environment from a predefined map; compute difference data representing a difference between the detected data and the reference data; and transfer the difference data, wherein an image computed based on the difference data and the reference data represents the detected data.